another of those "duh! what was I thinking" moments, this specific one
being "why test that files/directories are created with the right user
and group IDs? Shouldn't that be out of your control, as well as
totally unnecessary on a sane system?"
The initial fetch of a new repo which has 'upstream' read-only mirroring
configured will cause a spurious error concerning FETCH_HEAD not yet
existing. This silences the error.
while we're about it, add the same check to some of the internal
routines, so that commands can also be protected.
finally, just to make sure we don't lose it again in some other fashion,
add a few tests for path traversal...
before this, trying to access a wild repo would create an empty hash in
%repos. This is pretty harmless, but at some later point, memberships()
would try to use that in a pattern, attempting to match the real repo
being access-checked.
Which is still fine if your repo doesn't look like "libstdc++" AND
you're using some recent perl.
However, for perl 5.8.8, and if the repo has a ++ in it, perl barfs.
Here's a test program to check your perl:
#!/usr/bin/perl
$base="foo/u1/libstdc++";
$i="foo/u1/libstdc++";
if ( $base =~ /^$i$/ ) {
print 1;
} else {
print 2;
}
On 5.14.2 I get "2". On 5.8.8 I get:
Nested quantifiers in regex; marked by <-- HERE in m/^foo/u1/libstdc++ <-- HERE $/ at ./aa.pl line 6.
gitolite does indeed try to not load itself twice, but I forgot that by
that time the pwd is ~/.gitolite/conf not ~/.gitolite so it always ended
up reading itself twice in case of a wildcard include.
Add a special variable `%GL_CREATOR` to the the git-config trigger that
is replaced by the name of the repository creator (if any).
This can be useful to set up the default owner configuration for wild
repositories.
Example:
repo assignments/CREATOR/a[0-9][0-9]
C = @students
RW+ = CREATOR
config gitweb.owner = %GL_CREATOR
----
committer added an if condition to the s/// line.
Background
* I needed to have the documentation under CC-BY-NC-SA (something
happened to force me to choose)
* Distros don't like the NC part. They'd rather drop the
documentation entirely instead
* I don't like that; it bothers me that even a clueful guy won't be
able to do a basic install with what comes in the package.
* Meanwhile, I have always had the occasional "TL;DR" complaint about
my docs
Taking all this into account, it seemed like the best way was:
* Create a brand new README.txt that is crisp enough for someone to
glance through and quickly get started. At more then 300 lines, it
covers enough ground that probably 60% of sites don't need more.
Put this under the CC-BY-SA license, which is on the "good" list for
Fedora (and also Debian, I am told).
* Move the current documents to a new "gitolite-doc" repo that distros
can simply ignore, but anyone who has trouble can go to.
Make sure the online pages have the same content at the same URLs as
they do now, getting it instead from this new repo.
Link to the main URL in the new README.txt
- non-core documentation reduced to be easier to maintain
- much reduced progit section submitted to scott chacon, necessitating
some changes to this copy
- other minor stuff
- the "idiot-proof setup" :)
(plus get rid of that silly "dot.pl"; it's not needed any more, if it
ever was!)
Only '+' sign was unescaped in `http_simulate_ssh_connection()`.
When user translates `ssh git@host perms <repo> + <role> <user>` to
`curl https://host/git/perms?<repo>+%2b+<role>+<user>` nothing happens.
This commit fixes it modifying url unescaping.
committer notes: changed the regex per 'man URI::Escape'
(manually tested)
- new rc var: GL_BINDIR2; see doc update in this commit
- added _which() function to search both $GL_BINDIR and $GL_BINDIR2
- 'gitolite <command>', non-perl triggers, VREFs, and sugar, use this
- unshifted $GL_BINDIR2/lib into @INC upfront in Rc.pm
- perl triggers use this
- git version bumps up, causes minor change in t/merge-check.t
- for some strange reason apache cannot see /tmp/gitolite-http-authuserfile
(I thought private /tmp was only if you enabled selinux...)
